Shock Index and Postpartum Hemorrhage in Vaginal Deliveries: A Multicenter Retrospective Study

Abstract

Introduction: 

Shock index (SI), calculated by dividing heart rate by systolic blood pressure, is used to detect hemodynamic instability and hypovolemia. In obstetric practice, limited evidence is available regarding its usefulness in detecting postpartum hemorrhage (PPH). We aimed to evaluate the usefulness of SI in detecting PPH in vaginal deliveries using clinical data from 12 primary maternity care units in Japan.

Material and methods: 

In this multicenter retrospective study, a total of 30,820 women who delivered vaginally at term at 12 primary maternity care units from January 2012 to December 2018 were included. Systolic and diastolic blood pressures and heart rate were measured at five different time points from admission to postpartum 2 h, and postpartum blood loss was measured. We evaluated the trend of average SI and the performance of each vital sign for detection of PPH.

Results: 

The trend of average SI during labor and the immediate postpartum period was approximately 0.7 in women with blood loss of <500 mL. SI from the time of delivery of the placenta increased with an increase in blood loss. SI had the highest area under the receiver operating characteristic curve of 0.699 [95% confidence interval (CI), 0.682–0.716] and 0.758 (95% CI, 0.729–0.788) for PPH of ≥1,000 and ≥1,500 mL, respectively. However, both sensitivity of SI (1.0) for PPH (≥1,000 mL; 29.9%, and ≥1,500 mL; 40.5%, respectively) and correlation between maximum SI and blood loss (coefficient of correlation, 0.263) were low.

Conclusions: 

SI is a better parameter for PPH detection in vaginal deliveries than other vital signs. However, clinical judgment must incorporate other vital signs and symptoms associated with hypovolemic shock due to the low sensitivity of SI.

INTRODUCTION

Postpartum hemorrhage (PPH) affects approximately 5% to 6% of all pregnancies and is the principal cause of approximately one-quarter of all maternal deaths worldwide ^{(1, 2)}. Severe complications such as organ dysfunction and death can occur due to hypovolemic shock and coagulopathy from substantial blood loss ⁽³⁾. Majority of PPH-associated complications could be prevented by appropriate treatment; however, its timely recognition and diagnosis remain challenging for clinicians because impending shock may be masked by hemodynamic changes specific to pregnancy ^{(4, 5)}. Additionally, most PPH cases are diagnosed by clinical estimation of bleeding; however, studies have shown that blood loss estimation is not accurate, which may contribute to a delay in diagnosis and treatment delay ^{(6, 7)}.

The shock index (SI), calculated by dividing heart rate by systolic blood pressure (SBP), is used to detect hemodynamic instability and hypovolemia primarily in emergency medicine ⁽⁸⁾. It has also been widely used in obstetrics for estimating blood loss and as an indicator to initiate treatment for PPH ⁽⁹⁾. Recent studies have addressed its utility in predicting adverse maternal outcomes including intensive care unit admission, need for transfusion, and death among women with PPH compared with traditional vital signs such as heart rate and blood pressure ^{(10, 11)}. However, these are retrospective studies reporting only women diagnosed with PPH; thus, limited evidence is available regarding the usefulness of SI to detect PPH in the general obstetric population including non-PPH women. In addition, some guidelines emphasize using SI to detect PPH and evaluate its severity or the efficacy of its management ^{(12, 13)}; however, these guidelines do not provide a specific value of SI in PPH detection.

Many questions regarding SI during vaginal delivery remain unanswered ⁽¹⁴⁾: trend of SI during labor and the immediate postpartum period, whether SI is the optimal parameter to detect PPH, degree of correlation between SI and blood loss, how much blood loss is equivalent to an SI of 1.0, and characteristics (e.g., hypertensive disorders of pregnancy, obesity, and maternal age ≥35 years) of women who are likely to underestimate blood loss estimated by SI compared with actual blood loss. We aimed to answer these clinical questions to understand how SI might facilitate early recognition and treatment of PPH and to avoid pitfalls associated with SI.

PATIENT AND METHODS

We performed this multicenter retrospective study based on clinical data from 12 primary healthcare units for perinatal care located in Aichi and Gifu prefectures, Japan. All of these units care for women with low-risk pregnancies, with each unit conducting 300 to 1,000 deliveries annually. Data on women who delivered vaginally at these units from January 2012 to December 2018 were analyzed. Exclusion criteria were as follows: preterm delivery (<37 gestational weeks), post-term delivery (≥42 gestational weeks), cesarean section, multiple pregnancy, stillbirth, major congenital or chromosomal abnormalities, transfer to a higher-level facility, and missing data on vital signs and blood loss. Flow diagram of the study population is shown in Figure S1, https://links.lww.com/SHK/B102. Among 45,058 women who delivered after 22 gestational weeks at these units, a total of 30,820 were eligible and 14,238 women were excluded (Figure S1, https://links.lww.com/SHK/B102). Only women with complete data on vital signs and blood loss from admission to the units to postpartum 2 h were eligible to improve data reliability. All of the 12 units comprised a medical group; the basic management for delivery including data collection, vital sign measurement, blood loss estimation (gravimetric method), and management for the prevention of PPH at the third stage of labor was standardized among all 12 units.

Maternal demographic characteristics included age, gestational week at delivery, parity, prepregnancy body mass index (BMI), BMI at delivery, and use of any assisted reproductive technology. Obstetric and neonatal information collected included epidural anesthesia, labor induction, instrumental delivery, perineal laceration, hypertensive disorders of pregnancy, endometriosis, blood loss, infant sex, and birthweight. Clinical information was extracted from the medical records of each unit and cumulated for further analysis.

SBP, diastolic blood pressure (DBP), and heart rate were measured using an automated sphygmomanometer with a double cuff (Terumo ES-H55, Terumo Corporation, Tokyo, Japan) ⁽¹⁵⁾. SI, mean arterial pressure (MAP), and pulse pressure (PP) were calculated from these values. Vital signs were measured at five different time points: admission to a unit, at the first stage of labor, following delivery of the placenta, postpartum 1 h, and postpartum 2 h. Maximum SI was defined as the highest value among three time points (following delivery of the placenta, postpartum 1 h, and postpartum 2 h). Plausible ranges were defined as SBP of 50 to 220 mm Hg, DBP of 30 to 140 mm Hg, heart rate of 40 to 200 bpm, and SBP > DBP. Outliers were excluded from further analyses.

PPH was defined as a cumulative blood loss of ≥1,000 mL for vaginal deliveries ⁽¹⁶⁾. Quantification of blood loss was performed by the gravimetric method at three different time points: following delivery of the placenta, postpartum 1 h, and postpartum 2 h. Total blood loss did not include amniotic fluid and was measured by weighing pre- and post-procedure gauze and pads. Hypertensive disorders of pregnancy were defined as SBP of ≥140 mm Hg and/or DBP of ≥90 mm Hg during pregnancy ⁽¹⁷⁾.

Active management for the prevention of PPH at the third stage of labor, such as administration of prophylactic uterotonics (oxytocin) before delivery of the placenta and controlled cord traction, was routinely performed ^{(18, 19)}. Treatment for PPH includes the use of additional uterotonics (prostaglandins or ergometrine), uterine massage, and fluid resuscitation with isotonic crystalloids. Conservative interventions such as tranexamic acid use and intrauterine balloon tamponade were considered in cases of refractory atonic bleeding. However, transfusion and surgical interventions such as uterine artery embolization and hysterectomy were not available at the study units. Women requiring these interventions were referred to a higher-level health facility.

To provide reference values for SI during labor and the immediate postpartum period among women without pregnancy and obstetric complications, we excluded women with the following characteristics: prepregnancy BMI <18.5 kg/m² or ≥25 kg/m², endometriosis, hypertensive disorders of pregnancy, epidural anesthesia, induction of labor, instrumental delivery, third- or fourth-degree perineal laceration, prolonged labor, and blood loss ≥500 mL. The mean ± SD, 2.5th, 50th, and 97.5th percentiles for SI were calculated for the remaining 11,957 women.

We analyzed receiver operator characteristic curves to determine the optimal vital sign cutoff point to detect PPH during postpartum first 2 h. Area under the receiver operating characteristic curve (AUROC) for each parameter was calculated. Sensitivity, specificity, and positive and negative predictive values with 95% confidence intervals (CI) were determined for optimal detection. To eliminate the effect of body size, which was associated with a change in vital sign due to PPH ⁽²⁰⁾, we additionally analyzed the AUROC for each parameter stratified by the BMI category at delivery (<25 kg/m², 25–30 kg/m², and ≥30 kg/m²).

To determine the amount of blood loss equivalent to an SI of 1.0, we have provided a formula for the total blood loss using the fitted curve in a scatter plot (maximum SI and total blood loss) across various groups (all women, women with blood loss of ≥500 mL, and women with blood loss of ≥1,000 mL).

In the clinical scenario, we sometimes experience a discrepancy between the blood loss and SI. Thus, to identify obstetrical factors associated with a wide gap (≥500 mL) between actual blood loss and estimated blood loss using this formula, we performed univariate and multivariate analyses among women with blood loss of ≥500 mL. Crude and adjusted odds ratios (ORs) with 95% CI were evaluated in logistic regression models after adjustment for covariates including ≥35 years of age, primipara, prepregnancy body mass index of ≥25 kg/m², assisted reproductive technology, epidural anesthesia, labor induction, instrumental delivery, third- or fourth-degree perineal laceration, hypertensive disorders of pregnancy, blood loss, and birthweight. All statistical analyses were performed with SPSS 26 (SPSS Inc, Chicago, Ill) and JMP Pro 14 (SAS Institute Inc, Cary, NC). A P value of <0.05 was considered statistically significant.

This study protocol was approved by the Institutional Ethics Board of Nagoya University (approval number 2015-0415), and data use was authorized by the Kishokai Medical Corporation Executive Committee (approval number 2016-009). The need for signed informed consent was waived because data were anonymized and retrospectively collected from the existing medical records.

RESULTS

A total of 45,058 women delivered at ≥22 gestational weeks at the 12 primary maternity care units from January 2012 to December 2018. Among these, 14,238 women met the exclusion criteria, resulting in a total of 30,820 women with complete data who formed the study population (Figure S1, https://links.lww.com/SHK/B102). Sociodemographic and obstetric characteristics of this population are shown in Table 1. Total blood loss, without amniotic fluid, in postpartum 2 h was 285 mL (median). A total of 24,955 women had blood loss of <500 mL, 4,748 had blood loss between 500 and 999 mL, 823 had blood loss between 1,000 and 1,499 mL, and 294 had blood loss of ≥1,500 mL.

Table 1 - Sociodemographic and obstetric characteristics

Characteristics	n = 30,820
Maternal characteristics
Maternal age (years)	31 (28–34)
Gestational age at delivery (weeks)	39.3 (38.4–40.0)
Primipara (%)	55.2%
Prepregnancy BMI (kg/m2)	20.1 (18.8–21.9)
BMI at delivery (kg/m2)	24.5 (22.9–26.5)
Pregnancy via assisted reproductive technology (%)	5.1%
Epidural anesthesia (%)	1.8%
Induction of labor (%)	21.5%
Instrumental delivery (%)	10.3%
Third- or fourth-degree perineal laceration (%)	1.2%
Hypertensive disorders of pregnancy (%)	9.0%
Endometriosis (%)	0.6%
Blood loss (mL)	285 (192–437)
Neonatal characteristics
Male (%)	50.9%
Birthweight (g)	3,088 (2,858–3,332)

BMI indicates body mass index. Median (interquartile range) or %.

The trend of average SI from admission to postpartum 2 h, stratified by total amount of blood loss, is noted in Figure 1. The SI of women with a normal range of blood loss (<500 mL) remained essentially flat at approximately 0.7 throughout delivery, although SI increased briefly following delivery of the placenta. SI increased with an increase in blood loss following delivery of the placenta. Even in women with blood loss of ≥1,500 mL, the mean SI varied in the range of 0.80 to 0.85 and did not reach 0.9, which was previously reported to be significantly associated with maternal adverse outcomes ⁽¹⁰⁾. Table 2 shows the reference SI values among 11,957 women without pregnancy and obstetric complications. The mean ± SD, 2.5th, 50th, and 97.5th percentiles of SI at the different time points are presented in Table 2. The 97.5th percentile of SI following delivery was approximately 1.0.

Fig. 1: Trend of the average shock index from admission to postpartum 2 h stratified by total blood loss. Mean of the shock index from admission to postpartum 2 h was evaluated for all degrees of blood loss.

Table 2 - Reference values for shock index among 11,957 women without pregnancy and obstetric complications from admission to 2 h after birth

	Admission	1st stage of labor	After delivery of placenta	1 h after birth	2 h after birth
Mean ± SD	0.69 ± 0.13	0.68 ± 0.14	0.70 ± 0.14	0.67 ± 0.12	0.67 ± 0.13
2.5th percentile	0.47	0.45	0.47	0.46	0.45
50th percentile	0.67	0.67	0.68	0.65	0.66
97.5th percentile	1.00	1.02	1.01	0.94	0.97

SD indicates standard deviation.

Figure 2S, https://links.lww.com/SHK/B103 A–E notes the trends of SBP, DBP, MAP, heart rate, and PP from admission to postpartum 2 h stratified by total blood loss. There were no major changes in SBP, DBP, and MAP among the different categories of total blood loss (Figure 2S A–C, https://links.lww.com/SHK/B103). However, heart rate and PP changed considerably based on total blood loss (Figure 2S D and E, https://links.lww.com/SHK/B103).

To determine whether SI is better at detecting PPH than other vital signs, we evaluated the performance of vital signs based on AUROC. Table 3 summarizes the AUROC values with 95% CI for each vital sign to detect PPH (blood loss ≥1,000 mL and ≥1,500 mL). Regarding PPH (≥1,000 and ≥1,500 mL), SI had the highest AUROC of 0.699 (0.682–0.716) and 0.758 (0.729–0.788), respectively. This was significantly higher than the AUROC of heart rate (P < 0.01 and P < 0.01, respectively) and other vital signs (P < 0.01 and P < 0.01, respectively). The cutoff points of SI for PPH (≥1,000 and ≥1,500 mL) were 0.86 (sensitivity, 53.9%; specificity, 76.0%) and 0.83 (sensitivity, 69.7%; specificity, 69.4%), respectively. The cutoff points of heart rate for PPH (≥1,000 and ≥1,500 mL) were 91 bpm (sensitivity, 62.9%; specificity, 63.6%) and 102 bpm (sensitivity, 52.7%; specificity, 82.6%), respectively. Table S1, https://links.lww.com/SHK/B104 shows the performance of SI (0.70–1.5) for the detection of PPH of ≥1,000 and ≥1,500 mL as well as the prevalence of PPH. A combination of SI and heart rate had higher AUROC of 0.700 (0.684–0.717) and 0.761 (0.732–0.791), respectively, than those of SI alone; however, this did not reach statistical significance (P = 0.06 and P = 0.20, respectively). To eliminate the effect of body size on vital signs in PPH, we additionally evaluated the AUROC for each parameter stratified by BMI category at delivery (<25 kg/m², 25–30 kg/m², and ≥30 kg/m²). We found that SI was the highest AUROC irrespective of the BMI category at delivery (Table S2, https://links.lww.com/SHK/B104).

Table 3 - AUROC values with 95% confidence interval of vital signs parameters to detect postpartum hemorrhage

Vital sign	AUROC (95% CI) for PPH (≥1,000 mL)	AUROC (95% CI) for PPH (≥1,500 mL)
Shock index	0.699 (0.682–0.716)	0.758 (0.729–0.788)
Heart rate	0.679 (0.663–0.696)	0.733 (0.704–0.762)
Pulse pressure	0.634 (0.617–0.651)	0.648 (0.614–0.681)
Systolic blood pressure	0.583 (0.564–0.601)	0.604 (0.569–0.640)
Diastolic blood pressure	0.542 (0.524–0.559)	0.568 (0.533–0.602)
Mean arterial pressure	0.520 (0.502–0.538)	0.543 (0.507–0.579)

AUROC indicates area under the receiver operating characteristic curve; CI, confidence interval; PPH, postpartum hemorrhage.

Figure 2 shows a graph with the maximum SI plotted against total blood loss with a fitted curve for the entire patient population, depicting the distribution. The coefficient of determination (R²) was 0.069 and the coefficient of correlation was 0.263, indicating a low correlation between maximum SI and blood loss.

Fig. 2: Relationship between maximum shock index and total blood loss. Maximum shock index versus total blood loss with a fitted curve among the 30,820 women.

In an obstetric clinical situation, the SI is often used to estimate the total blood loss and severity of PPH empirically (e.g., SI of 1.0 ≈ blood loss of 1.5 L and SI of 1.5 ≈ blood loss of 2.5 L). To verify this and to determine the amount of blood loss equivalent to an SI of 1.0, we estimated blood loss using the fitted curve in a scatter plot across various blood loss groups (all women, women with blood loss of ≥500 mL, and women with blood loss of ≥1,000 mL). The formulas for blood loss estimated by SI for the different groups are as follows:$$\text{All}\text{women},\text{blood}\text{loss}(\text{mL})=504{(\text{<span class="ej-keyword" data-value="SI">SI</span>})}^2-442(\text{<span class="ej-keyword" data-value="SI">SI</span>})+391,{\text{R}}^2=0.069$$$$\text{Women}\text{with}\text{blood}\text{loss}\text{of}\ge 500\text{mL},\text{blood}\text{loss}(\text{mL})=73{(\text{<span class="ej-keyword" data-value="SI">SI</span>})}^2+240(\text{<span class="ej-keyword" data-value="SI">SI</span>})+548,{\text{R}}^2=0.049$$$$\text{Women}\text{with}\text{blood}\text{loss}\text{of}\ge 1,000\text{mL},\text{blood}\text{loss}(\text{mL})=-98{(\text{<span class="ej-keyword" data-value="SI">SI</span>})}^2+468(\text{<span class="ej-keyword" data-value="SI">SI</span>})+1033,{\text{R}}^2=0.026$$

We found that an SI of 1.0 is equivalent to a blood loss of 453 mL when all women were included, 861 mL exclusively for women with blood loss of ≥500 mL, and 1,403 mL for women with blood loss of ≥1,000 mL.

We sometimes experience a discrepancy between the blood loss and SI. Thus, to identify obstetric factors associated with a wide gap (≥500 mL) between actual blood loss and the blood loss estimated by SI using a fitted curve, we performed univariate and multivariate regression analyses for women with blood loss of ≥500 mL (Table 4):

Table 4 - Univariate and multivariate analyses of risk factors associated with a wide gap between blood loss estimated by the shock index and actual blood loss

Maternal and neonatal characteristics	Gap ≥500 mL (n = 476)	Gap <500 mL (n = 5,389)	Crude OR (95% CI)	Adjusted OR (95% CI)
Maternal age ≥ 35 years	149/476 (31.3)	1,244/5,389 (23.1)	1.52 (1.24–1.86)	1.73 (0.88–3.40)
Primipara (%)	290/476 (60.9)	3,411/5,389 (63.3)	0.90 (0.75–1.10)	1.02 (0.53–1.96)
Prepregnancy BMI (kg/m2) ≥ 25	48/474 (10.1)	525/5,367 (9.8)	1.04 (0.76–1.42)	0.30 (0.11–0.81)
Pregnancy via ART (%)	78/342 (22.8)	439/3,974 (11.1)	2.38 (1.81–3.12)	0.70 (0.31–1.56)
Epidural anesthesia (%)	14/476 (2.9)	122/5,389 (2.2)	1.31 (0.75–2.29)	2.49 (0.41–15.1)
Induction of labor (%)	199/476 (41.8)	1,634/5,389 (30.3)	1.65 (1.36–2.00)	1.49 (0.78–2.87)
Instrumental delivery (%)	106/473 (22.4)	914/5,357 (17.1)	1.40 (1.12–1.76)	1.01 (0.46–2.22)
Third- or fourth-degree perineal laceration (%)	23/461 (5.0)	124/5,239 (2.4)	2.17 (1.37–3.42)	0.71 (0.11–4.81)
Hypertensive disorders of pregnancy (%)	67/476 (14.1)	570/5,389 (10.6)	1.39 (1.06–1.82)	2.71 (1.15–6.39)
Blood loss (mL)	1,689 ± 385	725 ± 200
Birthweight (g)	3,334 ± 410	3,228 ± 360

Multivariate analysis was adjusted for covariates including maternal age of ≥35 years, primipara, prepregnancy BMI of ≥25 kg/m², assisted reproductive technology, epidural anesthesia, induction of labor, instrumental delivery, third- or fourth-degree perineal laceration, hypertensive disorders of pregnancy, blood loss, and birthweight.ART indicates assisted reproductive technology; BMI, body mass index; CI, confidence interval; OR, odds ratio.

Gap (mL) = actual blood loss (mL) − estimated blood loss [= 73 (SI)² + 240 (SI) + 548] (mL).

We found that women with hypertensive disorders of pregnancy were more likely to have blood loss underestimated by SI than actual blood loss (OR, 1.39; 95% CI, 1.06–1.82; adjusted OR, 2.71; 95% CI, 1.15–6.39).

DISCUSSION

In this multicenter retrospective study, we aimed to answer clinical questions related to SI in obstetric practice using clinical data from 12 primary maternity care units. We found that the trend of average SI during labor and the immediate postpartum period remained almost flat, at approximately 0.7, in women with blood loss of <500 mL and increased with an increase in blood loss following delivery of the placenta. SI was better at detecting PPH than other vital signs, and this result was consistent irrespective of BMI category at delivery. The combination of SI and heart rate was the most optimal for detecting PPH. SI had a high negative predictive value, but its sensitivity and positive predictive value were low. The correlation between maximum SI and blood loss was unexpectedly low. Although we were unable to determine the exact amount of blood loss equivalent to an SI of 1.0, our results indicate that this is estimated to be 1,500 mL in cases with severe PPH. Finally, SI was likely to underestimate blood loss in women with hypertensive disorders of pregnancy.

We found the trend of average SI to be approximately 0.7 during labor and the immediate postpartum period. This is consistent with findings of previous studies demonstrating that average SI values ranged from 0.68 to 0.74 within postpartum 2 h and were 0.66 (0.52–0.89; 90% reference range) within postpartum 1 h ^{(20, 21)}. Normal SI range has been reported to be 0.5 to 0.7 in the nonpregnant population, which is slightly lower than in the obstetric population ⁽⁸⁾. Given that the cutoff points of SI for detecting PPH (≥1,000 and ≥1,500 mL) were 0.86 and 0.83, respectively, we suggest that an SI of 0.85 to 0.90 may be an early sign of hypovolemia requiring appropriate intervention.

In the nonpregnant population, several reports have noted that an SI of 1.0 is equivalent to blood loss of 750 to 1,500 mL ⁽⁸⁾. However, the exact amount of blood loss equivalent to an SI of 1.0 in the obstetric population has not been well documented. We were unable to determine a specific amount of blood loss when SI was 1.0 because there were many false-positive cases of PPH, and the correlation between maximum SI and total blood loss was low. In our study, an SI of 1.0 may be equivalent to approximately 1,500 mL of blood loss under conditions of severe bleeding. Increased blood loss in the obstetric population compared with that in the nonpregnant population equating to an SI of 1.0 is primarily due to the systemic and physiological changes specific to pregnancy including increased circulatory blood volume, decreased systemic vascular resistance, and increased cardiac output ^{(14, 22)}.

The mechanism underlying the underestimation of blood loss in women with hypertensive disorders of pregnancy remains unclear. This might be explained by the fact that SI remains relatively lower due to higher SBP, even if heart rate increases with PPH. Another possible explanation is that women with hypertensive disorders of pregnancy are likely to have reduced intravascular volume and impairment of vascular reactivity and left ventricular function ⁽²³⁾. Therefore, we suggest that careful attention is needed in evaluating SI in women with hypertensive disorders of pregnancy even if SI is <1.0.

The high negative predictive value of SI indicates that it can be useful in ruling out PPH. However, of particular note in clinical practice is the low sensitivity of SI that will limit its ability, as well as that of other vital signs, to detect PPH. Therefore, estimated blood loss, other vital signs (heart rate and PP), and symptoms and examination findings associated with hypovolemic shock (pallor, prostration, perspiration, absent pulses, tachypnea, dizziness, confusion, cool and clammy skin, and capillary refill time > 2 s) should be considered for a provider's clinical judgment to accomplish timely recognition and diagnosis of PPH ⁽²⁴⁾. In addition, point-of-care coagulation tests, such as thromboelastography, must be taken into account in PPH management ⁽²⁵⁾. Simultaneously, it is important that initial treatment is not delayed by an obsession with accurate diagnosis and that overtreatment is acceptable in the management of PPH.

Among the strengths of this study is that it is the first, to the best of our knowledge, to evaluate various clinical questions associated with SI during labor and the immediate postpartum period. Although SI has been widely used and accepted in obstetric practice, evidence for its usefulness in detecting PPH in this setting is not well documented. Second, several prior studies examining SI included women who delivered by cesarean section ^{(10, 11)}, whereas we focused exclusively on women who delivered vaginally. Lastly, our sample size was larger than of prior studies ^{(10, 11, 21, 26)} and our study included data from multiple centers that improves the generalizability of our results.

There are several limitations that should be acknowledged. First, we were unable to consider several factors that can affect SI including total intravenous infusion volume, timing of various interventions, transfusion, and the use of vasopressors during postpartum 2 h. Vital signs can change dramatically in response to these therapeutic interventions, and information on these interventions was not included. However, none of these primary maternity care units have a blood bank on site, and women requiring transfusion, vasopressors, or surgical intervention were transferred to higher-level delivery facilities and were excluded from this study. Second, we only evaluated SI at three different time points following birth. Previously, the median time from PPH detection to the time of highest SI was found to be 15 min (4–32 min) ⁽¹⁰⁾; therefore, the maximum SI in this study may have been underestimated compared with the true maximum SI. However, Borovac-Pinheiro et al. ⁽²⁰⁾ demonstrated that the maximum SI occurred immediately following delivery of the placenta. Finally, as many previous reports have suggested, estimated blood loss by visual or gravimetric methods is not reliable and often underestimates blood loss ^{(6, 7)}. In addition, it is possible that a small amount of amniotic fluid was counted in total blood loss.

CONCLUSION

The trend of average SI during labor and the immediate postpartum period was approximately 0.7 in women with blood loss of <500 mL. We suggest that an SI of 0.85 to 0.90 is an early sign of hypovolemia and requires timely intervention. Although the sensitivity of SI is low, SI is a better parameter for PPH detection than other vital signs. Incorporating SI into PPH management may contribute to timely recognition of PPH and reduce blood loss. Estimated blood loss, vital signs, and symptoms of hypovolemic shock should be incorporated into clinical judgment to recognize and diagnose PPH.